[plug] "Hyper-Threading, suffers from a serious security flaw,"

Quintin Lette qlette at gmail.com
Sun May 15 09:34:58 WST 2005


all of our (dual) xeon servers have HT (and enabled)

(HT actually came out first on the Xeon IIRC then trickled over to the P4)

I also don't see us changing our stance on using HT in our terminal
servers as the cost of an additional server (which would be required
to sustain the load) is more than the risk of this vulnerability.

On 5/15/05, Craig Ringer <craig at postnewspapers.com.au> wrote:
> On Sat, 2005-05-14 at 22:50 +0800, Chris Watt wrote:
> > I thought the Xeon was a RISC processor (like the macinbtosh PowerPC)
> > which would mean it wouldn't need HT.
> 
> It's x86. Xeon is basically a P4 with more cache (sometimes waaay more
> cache) in a different form factor.
> 
> > > I've never seen a Xeon with HT..
> 
> I have two Xeons with HT.
> 
> > > I couldn't image anyone taking HT seriously for a server purpose...
> 
> Why not? For multi-user loads, it makes quite a significant difference
> to responsiveness, especially on servers with significant batch loads as
> well.
> 
> --
> Craig Ringer
> 
> _______________________________________________
> PLUG discussion list: plug at plug.org.au
> http://www.plug.org.au/mailman/listinfo/plug
> Committee e-mail: committee at plug.linux.org.au
>



More information about the plug mailing list